Optimizer iteration result insertion & extraction

[jjmmg]

When optimizing, I need to know the result of every iteration just when it's executed, to export it to another app, and where it is added to the optimization results to add new ones. The final result, when the optimization is finished, is unusefull for me. Is there a way to make it? I've implemented my own optimizer, and I noticed a mehod, RunIteration, where it could be done, but to overload it (if possible) I would need to know its code. If this is not a possibility, could someone tell me how I could achieve it?

[NinjaTrader_Bertrand]

jjmmg, this is unfortunately not supported, but you can create your own optimizer type where you store your iteration performance under RunIterationCompleted() / Iterate().

[jjmmg]

Bertrand, thanks for your prompt reply. I am creating my own optimization method, but I can't see anything about RunIterationCompleted method. I've search at http://www.ninjatrader-support.com/H....html?Overview, google, and different forums, but nothing about it.

It seems like it's exactly what I need. Where can I find a description of that method, or sample code, to look how and what kind of data is stored?

Correction: I've seen an "IterationCompleted" callback method in GeneticOptimizer. I guess it's what you meant, and what I need. I'll follow that way. Thanks.

[NinjaTrader_Bertrand]

You're welcome and correct, that's what I thought you would need to look into.

[jjmmg]

After having a look at it, I realized that this is the way to get the iteration performance to do whatever you want with it, but it's not what I meant. I suppose that NT stores the performance for the specific parameters set, and then passes it to the user through the callback method. Is there a way to acces to this performance before NT stores it, so I can modify it?

[NinjaTrader_Bertrand]

Unfortunately I'm not aware of a way to access this info before it's passed over.

[jjmmg]

And what about overloading RunIteration method, where I suppose it's done? Is there any unsupported documentation about the implementation of this method? Once this is done we could achieve the improvement we want to add, many computers running blocks of the same optimization. Doen't it sound interesting?

[NinjaTrader_Bertrand]

Jim, I agree it sounds like an interesting project, but unfortunately I would not have further info available.

[Trader.Jon]

jjmmg,

It occurs to me that the problem you are trying to solve might be something the programmers of the GO 's used in NT might be able to give some feedback to assist?!

Jon

[nikolaalx]

Has there been any progress on this?

I am also interested in creating an external module for parallel backtesting.

[NinjaTrader_PatrickH]

Hello nikolaalx,

I am not aware of any implementation of this, but I will follow up with development on this.

[NinjaTrader_PatrickH]

Hello nikolaalx,

Thank you for your patience.

There is no word on if or when this would be implemented.